Rheumatoid arthritis (RA) is the most common auto-immune arthritis with a prevalence between 0.5 - 2.5%. It is associated with substantial morbidity, mortality, and cost. Disease modifying anti-rheumatic drugs (DMARDs) reduce inflammation through a variety of mechanisms, improving short and long-term outcomes in RA. While there are a variety of drug-related toxicities, DMARDs are now accepted as standard of care for the medical management of RA, being recommended by the National Committee on Quality Assurance, the American College of Rheumatology (ACR), and the European League Against Rheumatism. Despite the recognized benefits of DMARDs, several published studies from the US, Canada, and Europe have found that 25-50% of patients diagnosed with RA do not use DMARDs, indicating sub-optimal care. There seem to be two major factors driving DMARD non-use: lack of access to expert rheumatic disease care and patient preferences. Access to rheumatologists is an acknowledged problem based on a report of the ACR finding that the aging US population will contribute to a 50% shortage of rheumatologists by 2025. It is very unlikely that rheumatology training programs can make up the deficit, suggesting that primary care providers (PCPs) will need to increase their familiarity with managing rheumatologic conditions, including RA. This trend coincides with the current push for primary care redesign, allowing for more comprehensive and collaborative management of diseases sometimes managed by specialists. An improved system for co-management between PCPs and rheumatologists around diseases like RA is needed. The first aim of this project, a 2x2 factorial randomized controlled trial, will test two different interventions. One will attempt to forge new collaborative care relationships between PCPs and rheumatologists, within the context of managing RA. The second will focus on patients and their preferences regarding RA treatment. Like other medications, patients express concerns about DMARD toxicities and how to balance risks and benefits. Addressing such issues appears to improve treatment compliance in other conditions and may help reduce DMARD non-use. We will attempt to address these issues through a telephonic health education program. The second aim will examine the economic implications of the intervention. We will use simulation models to calculate the long-term implications of the interventions tested in the aim one.